What Is the Source of Your Edge??

September 7th, 2020, 10:08 AM

TWDsje

Don't be silly. It should be easy for you to just create a version of the strategy using the Bollinger Bands and post the results. Show the full stats too because just seeing the cumulative net profit is not very interesting or useful.

Not being silly, just been there done that.

September 7th, 2020, 12:07 PM

TWDsje

This is where simply having a backtest that shows a profit is not sufficient to prove "edge". Any strategy that is going long over the past 10 years is probably going to have positive results. Plus your strategies are swing trading. Holding overnight provides some edge. So to prove that your technical edges are real, we need to look at risk adjusted measures. For instance, the Sharpe Ratio which in your case is .5.

We can also compare the backtest to similar variations. For instance, trading the opposite of the individual signal components, or replacing any use of a Donchian Channel with a Bollinger Band. And that's where I'm saying that in all of the research I have done, the Bollinger Band consistently outperforms other channel indicators. Then we can say (at least in that particular setup) that one indicator provides an edge over the other.

I have found it is extremely common for traders that have some positive results to not understand where their numbers are actually coming from. Many are just deriving edge from taking on excessive risk for instance. So if we're going to take these results as evidence of a technical edge, more work would be required.

Interesting that you prefer bollinger I do as well. Lately I only use bollinger and psar but maybe not in the normal way. To me bollinger is like a speedometer showing how fast or slow relative to the speed limit of the road. I can drive without it and realize that I am going too fast or slow but it is nice to have especially in a new city. There are videos of John Bollinger saying that people don't use bands "correctly". I'm not sure if that is true or not no way to know but he likely had contact with people using them not how he imagined. "Bollinger Bands answer a question: Are prices high or low on a relative basis? By definition price is high at the upper band and price is low at the lower band." That is similar to: am I driving over or under the speed limit relative to the speed limit on this road. It doesn't mean I can't go much faster or slower on the road.

Donchian channel is somewhat opposite and uses the high and low to show a range (vs standard deviation and average with bollinger) so its input is more sensitive to price "anomalies". Donchian channel is like answering the question how fast or slow am I going compared to the fastest/slowest person recently on the road. When it was created traders were phoning up the broker to make a trade and seeing quotes in a newspaper. Today there are HFT, many more emotional traders, more sophisticated price manipulation attempts, and crazy speed of news and data compared to the "good old days". It makes sense that something that averages out the noise is "better" for that environment. You could modify donchian channel by altering it to smooth out the high and low input which might make it works better for you.

The other indicator I like to use is PSAR and to me it is like showing if the road is turning right or left. Obviously if the road is turning right you better be steering right as well or you could drive off a cliff. Sure you could catch it at the correct time before the road turns left and have a nice shortcut but that is more dangerous than following the road reacting to it changing instead of predicting it. That kinda goes with your point about strategies working when the market is going one way up and not when trending down. If you can only turn left that works fine racing on most nascar tracks.

SpeculatorSeth · September 7th, 2020, 01:09 PM

three86

There are videos of John Bollinger saying that people don't use bands "correctly". I'm not sure if that is true or not no way to know but he likely had contact with people using them not how he imagined.

Well that brings us to an interesting question about what the best way to use such indicators are. To me the most simple BollingerBand strategy would be fade it as it breaks past the bands. Something like this:

Code

			
if(CrossAbove(High, MyBands.Upper, 1))
{
	EnterShort();
}
if(CrossBelow(Low, MyBands.Lower, 1))
{
	EnterLong();
}

bollinger

Which doesn't really work for a Donchian Channel. You could maybe say what if it's just touching the channel the line instead of crossing it:

Code

if(Low[0] == MyBands.Lower[0])
{
	EnterLong();
}
if(High[0] == MyBands.Upper[0])
{
	EnterShort();
}

donchian

Which isn't really quite the same thing. You'll get a lot more entries in the donchian channel one that way, but the overall result is a similar profit factor. Both come out to break even basically. At which point you'll get the same response that Bollinger gave. "Well you're just using it wrong." Which then begins the rabbit hole of how many different ways to use one indicator. People will claim to have success, but without the strategy in your hands or full results it's hard to know why it works. Is it really the indicator, or are they fooling us by using excessive risk?

Which brings us to this kind of stuff:

WoodyFox

Not being silly, just been there done that.

Great! Prove it.

It should be easy to just produce backtested summary pages for both. Then we could move on to how much of that strategy is actually edge from the indicator and how much is from just the holding period. Curious that we didn't get that isn't it?

September 7th, 2020, 01:59 PM

TWDsje

Well that brings us to an interesting question about what the best way to use such indicators are. To me the most simple BollingerBand strategy would be fade it as it breaks past the bands. Something like this:

Code

			
if(CrossAbove(High, MyBands.Upper, 1))
{
	EnterShort();
}
if(CrossBelow(Low, MyBands.Lower, 1))
{
	EnterLong();
}

bollinger

Which doesn't really work for a Donchian Channel. You could maybe say what if it's just touching the channel the line instead of crossing it:

Code

if(Low[0] == MyBands.Lower[0])
{
	EnterLong();
}
if(High[0] == MyBands.Upper[0])
{
	EnterShort();
}

donchian

Which isn't really quite the same thing. You'll get a lot more entries in the donchian channel one that way, but the overall result is a similar profit factor. Both come out to break even basically. At which point you'll get the same response that Bollinger gave. "Well you're just using it wrong." Which then begins the rabbit hole of how many different ways to use one indicator. People will claim to have success, but without the strategy in your hands or full results it's hard to know why it works. Is it really the indicator, or are they fooling us by using excessive risk?

Which brings us to this kind of stuff:

Great! Prove it.

It should be easy to just produce backtested summary pages for both. Then we could move on to how much of that strategy is actually edge from the indicator and how much is from just the holding period. Curious that we didn't get that isn't it?

You didnt add slippage or commission. With 40,000+ trades it will most likely make these strategies, as simple as they are, a losing proposition.

Robert

September 7th, 2020, 02:19 PM

TWDsje

Well that brings us to an interesting question about what the best way to use such indicators are. To me the most simple BollingerBand strategy would be fade it as it breaks past the bands. Something like this:

Code

			
if(CrossAbove(High, MyBands.Upper, 1))
{
	EnterShort();
}
if(CrossBelow(Low, MyBands.Lower, 1))
{
	EnterLong();
}

bollinger

Great! Prove it.

It should be easy to just produce backtested summary pages for both. Then we could move on to how much of that strategy is actually edge from the indicator and how much is from just the holding period. Curious that we didn't get that isn't it?

First off...I said that edges are not all that hard to find. What I didn't say is there laying for the pickings. Its not quite that easy

---- Like @Silver Dragon said...Always include commission and slippage. You did neither and would cost you over $200,000.00 just in commissions. (Now your strategy is in the hole whether its Bollinger or Donchian.)

---- Your over trading. Throw in some study...Whats the market volatility like. Are the bands expanding or contracting? Whats their Slope doing? Compare it to the norm over a period of time. Are there any time cycles within price action that are consistent.
Do these things and you might find that Donchian gives a better (Simple) picture of what price action is trying to do.

---- You say bollinger bands are better? Fine, I challenge you to show me a reasonable backtest that a least shows profit. (My Donchian Strategy showed real consistent profit which accounted for commissions and slippage)

September 7th, 2020, 02:20 PM

I have used bollinger quite a few different ways and there are also many ways to configure it you can use shorter or longer period different or multiple standard deviations. You can tune it to one ticker and different for another or different time frames. I have done all of that for many years with lots of indicators several I have created myself some based or combinations of existing ones. I've realized that I don't need all that maybe because I have traded long enough to have a feel. Also because I don't want it and I enjoy entering trades and moving stops. Trading isn't just to make money for me. Same as I wouldn't want a self driving or even modern car because I enjoy driving and upgrading/fixing cars and everything that goes into it.

For me everything is about my risk. Psar shows me the risk or the edge of the road. For example to me the russell is going down since August 20th my risk per contract is currently $2300 because the current price 1545.5 is that far away from 1591.5 in dollars which is daily psar. I can fade any rise I want against that but I would NEVER EVER go long. I don't care too much about entry and concentrate more on managing the exit but I will rarely enter short if it isn't at the top of bollinger or just coming off the top of my "short term" 10min chart. However, I won't take entry every time that happens nor do I have a set exit. Exit to me is more like surfing I ride it until I am tired or the wave ends there is always another. If I could I would backtest this I have been programming since the 70's but more of it is discretionary than isn't. I could create well defined rules and then make other rules to detect when first rules stop working. I'm not against that it just isn't for me anymore.

September 7th, 2020, 02:45 PM

Probability and repeatability.

Psychology has been a tricky one, while trading has made me a lot more aware of myself than anything else in my life so far, its never a straight line. It changes quite a lot. However, it is possible to build confidance over time when you have above two in your corner.
My system is also exceedingly common these days (Most of it learned from future.io), attached image just for fun

My answer is not profound or full of technical knowledge, its just my 2c.

chartCapture

September 7th, 2020, 04:21 PM

TWDsje

Holding overnight provides some edge.

Pardon my ignorance but could you please expand upon this? Not trying to sidetrack the discussion just very curious as to what kind of edge can be attributed to holding period in 6E.

September 7th, 2020, 05:20 PM

TWDsje

Stop playing games dude. I can use the same tricks.

I've never seen someone spend so much effort on trying to avoid addressing such a simple thing. It could have just been a simple picture "ahh well in this case it didn't make a difference". Instead you play games so we all know you're just trying to bullshit and mislead people.

Moderator Notice

@TWDsje

Up until you said another member is "just trying to bullshit and mislead people," this discussion was going along fine, because it was on the merits of different approaches. It is perfectly fine to disagree with another member on the merits, and to be passionate about it, too.

It is not all right to make it personal and to make personal accusations.

The post quoted above has been deleted.

I am aware that you have made substantial and worthwhile contributions to the forum, and I hope you will continue to. But this is also not the first time you have been cautioned about making a disagreement personal. Members who post here must keep their tone and their expressions professional and at least minimally courteous to others.

I realize that you care deeply about trading, as do we all. But if you wish to continue here, you will need to maintain a professional, non-personal attitude and expression.

You may take this as a warning. An additional case of personal rudeness will result in at least a temporary ban.

Bob.

SpeculatorSeth · September 7th, 2020, 07:02 PM

bobwest

Moderator Notice

@TWDsje

Up until you said another member is "just trying to bullshit and mislead people," this discussion was going along fine, because it was on the merits of different approaches. It is perfectly fine to disagree with another member on the merits, and to be passionate about it, too.

It is not all right to make it personal and to make personal accusations.

The post quoted above has been deleted.

I am aware that you have made substantial and worthwhile contributions to the forum, and I hope you will continue to. But this is also not the first time you have been cautioned about making a disagreement personal. Members who post here must keep their tone and their expressions professional and at least minimally courteous to others.

I realize that you care deeply about trading, as do we all. But if you wish to continue here, you will need to maintain a professional, non-personal attitude and expression.

You may take this as a warning. An additional case of personal rudeness will result in at least a temporary ban.

Bob.

That's ok. I don't have to post about it.

What Is the Source of Your Edge??

Discussion in Psychology and Money Management

What Is the Source of Your Edge??